Scholarlink

<p>Microarray technique deceives with a tiny chip carrying thousands of genetic instructions. Microarray DNA technique allowed simultaneous gene evaluation. Pattern recognition methods are typically employed in feature selection data to compare health and cancer patient sample data. Multidimensional gene expression comprises mismatched, noisy, and redundant genes. This has been the biggest challenge for machine learning techniques. Because it disrupts the testing and training process and affects the effectiveness of the classification. To overcome these deficiencies, the gene (or) feature is very important. In this article, we use Grey Wolf Optimization (GWO) technique for choosing feature subset selection in high-dimensional microarray cancer data. The cardinality of feature subsets and the distinguishing capacity of such selected subsets are two competing objectives which are needed to be addressed to replicate the objective functions. Designed an objective function which is effective addresses these issues and considers five benchmark datasets for evaluating the proposed system’s efficiency. The results determine the proposed system efficiency, which can be traced out the gene space search and are also ready to locate better gene selection. The proposed approach is reporting a better classification accuracy score than other existing approaches like GA, NSGA, PSO, etc. The score was 96.55% for the Naive Bayes classifier and 100% for other classifiers on the Colon dataset. For the Leukemia dataset, 97.35% is the highest classification accuracy. Similarly, 100% for Lymphoma data, 93.33% for the WDBC dataset, and 93.48% for the DLBCL dataset.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Classification decision model in high dimensional microarray cancer data using Grey Wolf optimizer by feature subset selection

  • 期刊论文

  • 发表日期：

    2026年4月13日

查看全文
  • Swetha Dhamercherla,
  • Damodar Reddy Edla,
  • Updesh Kumar Jaiswal,
  • Suresh Dara,
  • Punit Gupta,
  • Ankit Vidyarthi

摘要

Microarray technique deceives with a tiny chip carrying thousands of genetic instructions. Microarray DNA technique allowed simultaneous gene evaluation. Pattern recognition methods are typically employed in feature selection data to compare health and cancer patient sample data. Multidimensional gene expression comprises mismatched, noisy, and redundant genes. This has been the biggest challenge for machine learning techniques. Because it disrupts the testing and training process and affects the effectiveness of the classification. To overcome these deficiencies, the gene (or) feature is very important. In this article, we use Grey Wolf Optimization (GWO) technique for choosing feature subset selection in high-dimensional microarray cancer data. The cardinality of feature subsets and the distinguishing capacity of such selected subsets are two competing objectives which are needed to be addressed to replicate the objective functions. Designed an objective function which is effective addresses these issues and considers five benchmark datasets for evaluating the proposed system’s efficiency. The results determine the proposed system efficiency, which can be traced out the gene space search and are also ready to locate better gene selection. The proposed approach is reporting a better classification accuracy score than other existing approaches like GA, NSGA, PSO, etc. The score was 96.55% for the Naive Bayes classifier and 100% for other classifiers on the Colon dataset. For the Leukemia dataset, 97.35% is the highest classification accuracy. Similarly, 100% for Lymphoma data, 93.33% for the WDBC dataset, and 93.48% for the DLBCL dataset.